ac多模式匹配 java_ac自动机实现多模式匹配

用到了java中的hashMap，可以节约内存。

支持中文，可直接添加使用。

代码的注释还比较详细：

/**

*

* @author hit_fantasy

*/

import java.util.*;

public class TrieCheck {

//根节点root

private TrieNode root;

public TrieCheck() {

root = new TrieNode();

}

//测试

public static void main(String[] args) {

TrieCheck test = new TrieCheck();

test.addWord("shr");

test.addWord("say");

test.addWord("her");

test.addWord("he");

test.addWord("?");

test.buildFail(test.root);

boolean haha = test.hasBadWord("shrhaha?heabasd");

}

//添加模式串

public void addWord(String word) {

if (word == null || word.length() == 0) {

return;

}

TrieNode current = root;

for (int i = 0; i < word.length(); i++) {

char code = word.charAt(i);

current = current.add(code);

}

current.end = true;

current.word = word;

}

//查找过程

public boolean hasBadWord(String text) {

if (text == null || text.length() == 0 || root == null) {

return false;

}

TrieNode current = root;

TrieNode next;

boolean res = false;

for (int i = 0; i < text.length(); i++) {

next = current.child.get(text.charAt(i));

while (next == null) {//失配，转向失效函数

//转到自己对应的失效节点

current = current.fail.get(current);

if (current.end) {

res = true;

System.out.println(current.word);

}

//尝试继续往下走，直到能走通(next != null)

next = current.child.get(text.charAt(i));

//或者到了根节点

if (current == root && next == null) {

break;

}

}

if (next != null) {

current = next;

}

if (current.end) {

res = true;

System.out.println(current.word);

}

}

return res;

}

//建立失效函数

public void buildFail(TrieNode root) {

//队列queue在构建失效函数时使用

Queue queue = new LinkedList();

//root节点特殊处理

TrieNode current, next;

root.fail.put(root, root);

//根节点和根的孩子节点特殊处理，指向root

TrieNode father = root;

for (Map.Entry entry : root.child.entrySet()) {

entry.getValue().fail.put(entry.getValue(), father);

queue.add(entry.getValue());

}

while (!queue.isEmpty()) {

father = queue.remove();

for (Map.Entry entry : father.child.entrySet()) {

//当前节点父状态的失效节点,输入当前字符能走到哪

TrieNode tempFather = father;

current = tempFather.fail.get(tempFather).child.get(entry.getKey());

while (current == null) {

tempFather = tempFather.fail.get(tempFather);

current = tempFather.fail.get(tempFather).child.get(entry.getKey());

if (tempFather == root && current == null) {

entry.getValue().fail.put(entry.getValue(), root);

break;

}

}

if (current != null) {

entry.getValue().fail.put(entry.getValue(), current);

}

queue.add(entry.getValue());

}

}

}

//Trie节点数据结构

private static class TrieNode {

//当前节点的子节点

public Map child;

//当前节点的失效节点

public Map fail;

//标识当前节点是否为一个词的结束字符

private boolean end;

//记录一个词，输出时使用

public String word;

public char value;

public TrieNode() {

child = new HashMap();

fail = new HashMap();

end = false;

word = "";

}

//构建Trie树过程

public TrieNode add(char newChar) {

TrieNode t = child.get(newChar);

if (t == null) {

t = new TrieNode();

t.value = newChar;

child.put(newChar, t);

}

return t;

}

}

}